Jobs That AI Can’t Replace: The Impact of Automation on Workforce

First of all, there is a big difference between AI augmenting our workflows and completely replacing them. Currently, AI mostly nudge the decision-making vector in the right direction, which results in shorter time-to-action and higher decision accuracy.

For example, computer vision applications enable magnetic resonance imaging (MRI) tools to detect issues in soft tissues hundreds of times faster than human professionals. In this case, AI doesn’t make the surgeon’s skills obsolete but helps medics achieve the same results more efficiently. In finance, AI can simulate a market situation for humans to test different scenarios. Being a naturally more accurate mechanism than a human brain, it reduces error risks and cuts the time spent on analysis, insight generation, and taking action.

Such examples are prevalent across the majority of industries that are currently in the midst of AI disruption. Sectors like finance or healthcare are simply getting enhanced by streamlined AI-supported data analysis. In such cases, workers will only need to have a mild additional training to onboard new AI tools without losing their jobs. However, as AI will get more mature, it will have a much more significant impact on workforce depending on the nature of a job in question.

Most products we can’t imagine our lives without require hours of workers routinely performing the same set of tasks. In this case, regardless of human expertise, AI-enabled robots will achieve better results faster, with fewer resources. This is why jobs in sectors like manufacturing and transportation are the most threatened by automation in the long term. This includes crane operators, assemblers, service laborers, inspectional professionals, machinists, and electrical technicians.

According to the McKinsey research titled The Future of Mobility Is at Our Doorstep, autonomous vehicles will make up 40% of new vehicle sales in 2040. This means that many tasks, including physical allocation of resources in construction or remote control of emergency vehicles, can be completely automated, thus redefining fleet management development as we know it.

In the recent EY report Rethinking the Oil and Gas Workforce in 2040, researchers claim that 60% of equipment operators in Canada’s oil and gas sector will most likely be replaced by AI-enabled robots.

According to IDTechEX, more than 100K autonomous forklifts and trucks will be sold by 2030. Considering that more than 95% of forklifts are operated by humans and the ROI is achieved within 18 months on average, those kinds of navigational skills will become mostly obsolete in about 20 years.

It’s worth noting that regardless of the AI advancement level, industries that rely on people-to-people interaction won’t ever be completely automated. For example, the expertise of education professionals is measured more by social intelligence and empathy than any technical skills. Sales and marketing sectors require a granular understanding of people’s background and outspokenness — features that AI naturally can’t excel at. Human contribution is fundamental for many industries to operate.

Essentially, in the wake of mass AI adoption, the race for scarce AI talent will accelerate. Putting the right amount of resources into retraining will most likely differentiate leaders from the rest in this race. Here are the steps companies should consider when preparing for AI transformation in terms of workforce adaptation. 

• Plan for the long term but set short-term objectives

This will help you understand how specialized your AI needs are. Given that AI adoption is a continuous process, your AI talent demands will vary over time. By having a clear implementation roadmap, you can estimate how many workers you need to be retrained and how many you’ll need to attract from outside. 

• Evaluate your workforce’s skillsets

Basically, you need to find out who can be upskilled or retrained among your employees. Undeniably, some projects require building sophisticated AI models from the ground up, which calls for hiring competent AI and data engineers. For other projects, helping your business users learn how to utilize pre-built AI tools can be sufficient. 

• Work out a change management strategy

Given that AI is likely to be the most impactful technology this generation of workers will ever experience, their reluctance to change is not surprising. This is why it’s important to involve business leaders and opinion makers as early as possible. 

The reasons for AI integration should be transparently stated and discussed with exactly those people who will be the most affected. Leaders should highlight career development opportunities and provide a clear vision for assisting displaced employees if that’s the case. Moreover, AI usually implies more collaboration between different departments. By encouraging teamwork early on, it’s possible to maximize the cross-functional chemistry of different departments and facilitate their AI onboarding. 

• Involve end users into AI development

Companies need to do their best to create intuitive and user-friendly tools while also demonstrating the reasoning logic behind AI decisions. This is especially important for such sensitive industries as healthcare and banking, where practitioners need to rest assured that machine learning challenges are resolved in the process. Instilling confidence comes down not only to employees’ emotional reassurance but to technically proving that tools they are going to use can be trusted.  

• Educate beyond upskilling 

Let’s take healthcare as an example. As life expectancy is increasing, healthcare has to deal with more and more patients having complex chronic problems that require a proactive approach. With a growing number of AI applications proving helpful in this regard, we will need more practitioners being knowledgeable about using them. Basically, by 2030 an average clinician will need to be well-versed not only in their standard medical areas but also in AI and machine learning. This calls for a complete rethinking of healthcare education, where areas like medicine and IT intersect. Creating centers of excellence or comprehensive training programs focused on AI-centric education should become a new standard in the next decade.